Technical note: Reconstructing missing surface aerosol elemental carbon data in long-term series with ensemble learning
<p>Ground-based measurements of elemental carbon (EC) – classified under thermal–optical methods and considered a surrogate for black carbon – are essential for assessing air quality and evaluating climate impacts. However, data gaps caused by technical challenges impede comprehensive analyses...
Saved in:
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Copernicus Publications
2025-07-01
|
| Series: | Atmospheric Chemistry and Physics |
| Online Access: | https://acp.copernicus.org/articles/25/7485/2025/acp-25-7485-2025.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | <p>Ground-based measurements of elemental carbon (EC) – classified under thermal–optical methods and considered a surrogate for black carbon – are essential for assessing air quality and evaluating climate impacts. However, data gaps caused by technical challenges impede comprehensive analyses of long-term trends. This study proposed an ensemble learning modeling method to address these challenges. The model used readily accessible ground observation air pollutant data as proxies for EC-related tracers, along with meteorological parameters, to enhance prediction accuracy. It integrated outputs from Gradient Boosting Regression Trees, eXtreme Gradient Boosting, and random forest models, combining them through ridge regression to produce robust predictions. We applied this approach to reconstruct hourly EC concentrations from 2013–2023 for four cities in eastern China, filling 45 %–79 % of missing data and improving prediction performance by 8 %–17 % compared to individual models. Over the 11-year period, EC exhibited an overall decline (<span class="inline-formula">−0.20</span> to <span class="inline-formula"><math xmlns="http://www.w3.org/1998/Math/MathML" id="M2" display="inline" overflow="scroll" dspmath="mathml"><mrow><mo>-</mo><mn mathvariant="normal">0.14</mn><mspace linebreak="nobreak" width="0.125em"/><mrow class="unit"><mi mathvariant="normal">µ</mi><mi mathvariant="normal">g</mi><mspace width="0.125em" linebreak="nobreak"/><msup><mi mathvariant="normal">m</mi><mrow><mo>-</mo><mn mathvariant="normal">3</mn></mrow></msup><mspace linebreak="nobreak" width="0.125em"/><msup><mi mathvariant="normal">a</mi><mrow><mo>-</mo><mn mathvariant="normal">1</mn></mrow></msup></mrow></mrow></math><span><svg:svg xmlns:svg="http://www.w3.org/2000/svg" width="83pt" height="15pt" class="svg-formula" dspmath="mathimg" md5hash="e777538ad90b5cbf79febc3bc057989f"><svg:image xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="acp-25-7485-2025-ie00001.svg" width="83pt" height="15pt" src="acp-25-7485-2025-ie00001.png"/></svg:svg></span></span>), with a more significant decrease from 2013–2020 (<span class="inline-formula">−0.24</span> to <span class="inline-formula"><math xmlns="http://www.w3.org/1998/Math/MathML" id="M4" display="inline" overflow="scroll" dspmath="mathml"><mrow><mo>-</mo><mn mathvariant="normal">0.15</mn><mspace width="0.125em" linebreak="nobreak"/><mrow class="unit"><mi mathvariant="normal">µ</mi><mi mathvariant="normal">g</mi><mspace width="0.125em" linebreak="nobreak"/><msup><mi mathvariant="normal">m</mi><mrow><mo>-</mo><mn mathvariant="normal">3</mn></mrow></msup><mspace width="0.125em" linebreak="nobreak"/><msup><mi mathvariant="normal">a</mi><mrow><mo>-</mo><mn mathvariant="normal">1</mn></mrow></msup></mrow></mrow></math><span><svg:svg xmlns:svg="http://www.w3.org/2000/svg" width="83pt" height="15pt" class="svg-formula" dspmath="mathimg" md5hash="b9826ca6f58b8b926e9a6b8eae361393"><svg:image xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="acp-25-7485-2025-ie00002.svg" width="83pt" height="15pt" src="acp-25-7485-2025-ie00002.png"/></svg:svg></span></span>). During this time, the average EC concentration in the four cities dropped from 3.26 to 1.59 <span class="inline-formula">µg m<sup>−3</sup></span>, followed by a noticeable slowdown in the rate of decline from 2020–2023 (<span class="inline-formula">−0.12</span> to <span class="inline-formula"><math xmlns="http://www.w3.org/1998/Math/MathML" id="M7" display="inline" overflow="scroll" dspmath="mathml"><mrow><mo>-</mo><mn mathvariant="normal">0.04</mn><mspace linebreak="nobreak" width="0.125em"/><mrow class="unit"><mi mathvariant="normal">µ</mi><mi mathvariant="normal">g</mi><mspace linebreak="nobreak" width="0.125em"/><msup><mi mathvariant="normal">m</mi><mrow><mo>-</mo><mn mathvariant="normal">3</mn></mrow></msup><mspace width="0.125em" linebreak="nobreak"/><msup><mi mathvariant="normal">a</mi><mrow><mo>-</mo><mn mathvariant="normal">1</mn></mrow></msup></mrow></mrow></math><span><svg:svg xmlns:svg="http://www.w3.org/2000/svg" width="83pt" height="15pt" class="svg-formula" dspmath="mathimg" md5hash="6dd325e45116bc98a76f93d2e16e5dfe"><svg:image xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="acp-25-7485-2025-ie00003.svg" width="83pt" height="15pt" src="acp-25-7485-2025-ie00003.png"/></svg:svg></span></span>). Additionally, a fixed emission approximation method based on ensemble learning was proposed to quantitatively analyze the drivers of long-term EC trends. The analysis revealed that anthropogenic emission controls were the predominant contributors, accounting for approximately 92 % of the changes in EC trends from 2013–2020. However, their influence weakened post-2020, contributing approximately 80 %. These findings highlighted that while China's Clean Air Actions implemented since 2013 have substantially reduced black carbon concentrations, sustained and enhanced strategies are still necessary to further mitigate black carbon pollution.</p> |
|---|---|
| ISSN: | 1680-7316 1680-7324 |